Update app.py

app.py

@@ -1316,17 +1316,13 @@ def bold_high_impact_row(row):
     except:
         pass
     return [''] * len(row)
-import gradio as gr
-import pandas as pd
-from itertools import zip_longest
-
 
-import gradio as gr
-import pandas as pd
-from itertools import zip_longest
 
 def variable_outputs(file_inputs):
-    # Run processing
+
+    file_inputs_single = file_inputs
+   
+    # Call batch file processing and get analysis results
     (
         img_bucketfull_list,
         img_trust_list,
@@ -1336,11 +1332,16 @@ def variable_outputs(file_inputs):
         img_satisfaction_list,
         df_builder_pivot_list,
         output_text_list,
-    ) = batch_file_processing(file_inputs)
+    ) = batch_file_processing(file_inputs_single)
+
+    # Get number of datasets uploaded
+    k = len(file_inputs_single)
 
-    k = len(file_inputs)
-    output_components = []
+    # Container for visible plots
+    global plots_visible
+    plots_visible = []
 
+    # Use zip_longest to iterate over the lists, padding with None
     for row, (
         img_bucketfull,
         img_trust,
@@ -1350,49 +1351,115 @@ def variable_outputs(file_inputs):
         img_satisfaction,
         df_builder_pivot,
         output_text,
-    ) in enumerate(zip_longest(
-        img_bucketfull_list,
-        img_trust_list,
-        img_nps_list,
-        img_loyalty_list,
-        img_consideration_list,
-        img_satisfaction_list,
-        df_builder_pivot_list,
-        output_text_list,
-    )):
-        dataset_name = file_inputs[row].split("/")[-1]
-
-        # Header and description
-        output_components.append(gr.Markdown("<span style='font-size:20px; font-weight:bold;'>Trust and NPS Drivers</span>", visible=True))
-        output_components.append(gr.Markdown("The analysis identifies the TrustLogic® dimensions that are most effective in driving your audience's likelihood to recommend and trust you.", visible=True))
-
-        # Plots
-        output_components += [
-            gr.Image(value=img_trust, type="pil", label="Trust Drivers", visible=img_trust is not None),
-            gr.Image(value=img_nps, type="pil", label="NPS Drivers", visible=img_nps is not None),
-            gr.Image(value=img_loyalty, type="pil", label="Loyalty Drivers", visible=img_loyalty is not None),
-            gr.Image(value=img_consideration, type="pil", label="Consideration Drivers", visible=img_consideration is not None),
-            gr.Image(value=img_satisfaction, type="pil", label="Satisfaction Drivers", visible=img_satisfaction is not None),
+    ) in enumerate(
+        zip_longest(
+            img_bucketfull_list,
+            img_trust_list,
+            img_nps_list,
+            img_loyalty_list,
+            img_consideration_list,
+            img_satisfaction_list,
+            df_builder_pivot_list,
+            output_text_list,
+        )
+    ):
+        # Get dataset name
+        dataset_name = file_inputs_single[row].split("/")[-1]
+        global plots
+        # Based on the number of files uploaded, determine the content of each textbox
+        plots = [
+            
+            gr.Markdown(
+                "<span style='font-size:20px; font-weight:bold;'>Trust and NPS Drivers</span>",
+                visible=True,
+            ),
+            gr.Markdown(
+                """
+                The analysis identifies the TrustLogic® dimensions that are most effective in driving your audience's likelihood to recommend and trust you
+                """,
+                visible=True,
+            ),
+
+            # ✅ Side-by-side Trust & NPS drivers
+            gr.Image(
+                value=combine_two_images_horizontally(img_trust, "Trust Drivers", img_nps, "NPS Drivers"),
+                type="pil",
+                label="Trust + NPS Drivers",
+                visible=True,
+            ),
+            gr.Image(
+                value=None,
+                type="pil",
+                visible=False,
+            ),
+
+
+            gr.Image(
+                value=None,
+                type="pil",
+                visible=False,
+            ),
+            gr.Image(
+                value=None,
+                type="pil",
+                visible=False,
+            ),
+            gr.Image(
+                value=None,
+                type="pil",
+                visible=False,
+            ),
+            gr.Textbox(
+                value=output_text,
+                visible=False,
+            ),
         ]
 
-        # Table section
+        # add current plots to container
+        plots_visible += plots
+
         if isinstance(df_builder_pivot, pd.DataFrame):
-            output_components.append(gr.Markdown("<span style='font-size:20px; font-weight:bold;'> What to say and do to build your trust and Net Promoter Score </span>", visible=True))
-            output_components.append(gr.Markdown("<span style='font-size:17px; font-weight:bold;'>The features and messages that improve your NPS and trust most effectively. The key ones are marked in bold.</span>", visible=True))
-            output_components.append(gr.Dataframe(value=df_builder_pivot, label=f"{dataset_name}", visible=True, height=800))
-        else:
-            output_components += [
-                gr.Markdown("", visible=False),
-                gr.Markdown("", visible=False),
-                gr.Dataframe(value=None, label="", visible=False),
-            ]
+            logger.debug(f"df_builder_pivot: {df_builder_pivot}")
+
+            markdown_5 = gr.Markdown(
+                "<span style='font-size:20px; font-weight:bold;'> What to say and do to build your trust and Net Promoter Score  </span>",
+                visible=True,
+            )
+
+            markdown_6 = gr.Markdown(
+                "<span style='font-size:17px; font-weight:bold;'>The  features and messages that improve your NPS and trust most effectively. They key ones are marked in bold.</span>",
+                #+ "<br>In the table, use the little arrow in each column to toggle the most to least effective TrustBuilders® to fill each Trust Bucket®. Your focus is only on the Trust Bucket® with the highest driver impact. "
+               # + "<br> Note: Even if Trust Buckets® for Customers and Prospects overlap, the most effective statements are very different. This provides clear guidance for acquisition versus loyalty activities.",
+                visible=True,
+            )
+
+            styled_df = df_builder_pivot.style.apply(bold_high_impact_row, axis=1)
 
-        # Additional notes if present
-        if isinstance(output_text, str) and output_text.strip():
-            output_components.append(gr.Markdown(f"**Additional Notes:**\n\n{output_text.strip()}", visible=True))
 
-    # Pad outputs to match expected component count if needed
+            table_builder_2 = gr.Dataframe(
+                value=styled_df,
+                headers=list(df_builder_pivot.columns),
+                interactive=False,
+                label=f"{dataset_name}",
+                visible=True,
+                height=800,
+                wrap=True,
+            )
+
+
+
+            plots_visible.append(markdown_5)
+            plots_visible.append(markdown_6)
+            plots_visible.append(table_builder_2)
+        else:
+            plots_visible.append(gr.Markdown("", visible=False))
+            plots_visible.append(gr.Markdown("", visible=False))
+            plots_visible.append(gr.Dataframe(value=None, label="", visible=False))
+
     plots_invisible = [
+        gr.Markdown("", visible=False),
+        gr.Markdown("", visible=False),
+        gr.Image(label="Trust Buckets", visible=False),
         gr.Markdown("", visible=False),
         gr.Markdown("", visible=False),
         gr.Image(label="Trust Drivers", visible=False),
@@ -1400,27 +1467,89 @@ def variable_outputs(file_inputs):
         gr.Image(label="Loyalty Drivers", visible=False),
         gr.Image(label="Consideration Drivers", visible=False),
         gr.Image(label="Satisfaction Drivers", visible=False),
+        gr.Textbox(label="Analysis Summary", visible=False),
         gr.Markdown("", visible=False),
         gr.Markdown("", visible=False),
         gr.Dataframe(value=None, label=" ", visible=False),
     ]
 
-    return output_components + plots_invisible * (max_outputs - k)
+    return plots_visible + plots_invisible * (max_outputs - k)
+
 
 def reset_outputs():
-    return [
-        gr.Markdown(visible=True),  # markdown_3
-        gr.Markdown(visible=True),  # markdown_4
-        gr.Image(visible=True),     # trust_plot
-        gr.Image(visible=True),     # nps_plot
-        gr.Image(visible=True),     # loyalty_plot
-        gr.Image(visible=True),     # consideration_plot
-        gr.Image(visible=True),     # satisfaction_plot
-        gr.Textbox(visible=True),   # summary_text
-        gr.Markdown(visible=True),  # markdown_5
-        gr.Markdown(visible=True),  # markdown_6
-        gr.Dataframe(visible=True)  # df_builder_pivot
-    ]
+    # Reset outputs
+    outputs = []
+
+    # Create fixed dummy components
+  
+
+    markdown_3 = gr.Markdown(
+        "<span style='font-size:20px; font-weight:bold;'>Trust and NPS Drivers</span>",
+        visible=True,
+    )
+    markdown_4 =  gr.Markdown(
+                """
+                This analysis shows which Trust Buckets® are most effective in building trust and improving your key performance indicators (KPIs).
+                <br><br>
+                The middle line is the average importance. The bars extending to the right show which Trust Buckets® are most important. The higher the percentage, the more important the Trust Bucket® is to your audience.
+                """,
+                visible=True,
+            )
+    trust_plot = gr.Image(value=None, label="Trust Drivers", visible=False)
+    nps_plot = gr.Image(value=None, label="NPS Drivers", visible=False)
+    loyalty_plot = gr.Image(value=None, label="Loyalty Drivers", visible=False)
+    consideration_plot = gr.Image(
+        value=None, label="Consideration Drivers", visible=False
+    )
+    satisfaction_plot = gr.Image(value=None, label="Satisfaction Drivers", visible=False)
+    summary_text = gr.Textbox(value=None, label="Analysis Summary", visible=False)
+
+    markdown_5 = gr.Markdown(
+        "<span style='font-size:20px; font-weight:bold;'>TrustBuilders®",
+        visible=True,
+    )
+    markdown_6 = gr.Markdown(
+    "These are the specific reasons to trust and recommend. They can be your brand values, features, attributes, programs, and messages. "
+    + "<br>For practical purposes, they tell you exactly what to do and say to build more trust and improve your KPIs. "
+    + "<br>In the table, use the little arrows to toggle by Trust Bucket® or Trust Builder® importance. "
+    + "<br>Tip: Compare Owners and Prospects. Even though some of the Trust Buckets® are the same, the Trust Builders® are very different.",
+      visible=True,
+    )
+
+
+    df_builder_pivot = gr.Dataframe(value=None, label="", visible=True)
+
+
+    outputs.append(markdown_3)
+    outputs.append(markdown_4)
+    outputs.append(trust_plot)
+    outputs.append(nps_plot)
+    outputs.append(loyalty_plot)
+    outputs.append(consideration_plot)
+    outputs.append(satisfaction_plot)
+    outputs.append(summary_text)
+    outputs.append(markdown_5)
+    outputs.append(markdown_6)
+    outputs.append(df_builder_pivot)
+
+    # invisible from second set onwards
+    for i in range(1, max_outputs):
+        outputs.append(gr.Markdown("", visible=False))
+        outputs.append(gr.Markdown("", visible=False))
+        outputs.append(gr.Image(value=None, label="", visible=False))
+        outputs.append(gr.Markdown("", visible=False))
+        outputs.append(gr.Markdown("", visible=False))
+        outputs.append(gr.Image(value=None, label="", visible=False))
+        outputs.append(gr.Image(value=None, label="", visible=False))
+        outputs.append(gr.Image(value=None, label="", visible=False))
+        outputs.append(gr.Image(value=None, label="", visible=False))
+        outputs.append(gr.Image(value=None, label="", visible=False))
+        outputs.append(gr.Textbox(value=None, label="", visible=False))
+        outputs.append(gr.Markdown("", visible=False))
+        outputs.append(gr.Markdown("", visible=False))
+        outputs.append(gr.Dataframe(value=None, label="", visible=False))
+
+    return outputs
 
 
 def data_processing(file_path):
@@ -2275,10 +2404,6 @@ def display_trust_score_2():
 from sklearn.linear_model import LinearRegression
 from sklearn.metrics import r2_score
 
-import base64
-
-
-
 import base64
 
 def reset_variables():
@@ -2427,200 +2552,158 @@ def load_nps_and_r2(file_path):
     return nps_img, r2_img
 
 
+
 with gr.Blocks(css=css, js=js_func) as demo:
-   gr.HTML("""
-   <style>
-   #trust_driver_img, #nps_driver_img {
-       display: inline-block;
-       width: 49%;
-       margin-right: 1%;
-       vertical-align: top;
-   }
-   </style>
-   """)
-
-   # 🔁 States
-   vw_customers_state11 = gr.State(value=["example_files/VW Owners.xlsx"])
-   vw_prospects_state12 = gr.State(value=["example_files/Volkswagen Non Customers.xlsx"])
-   vw_customers_state = gr.State(value=["VW Owners.xlsx"])  # Just the filename
-   vw_prospects_state = gr.State(value=["Volkswagen Non Customers.xlsx"])  # Just the filename
-   selected_dataset_ai = gr.State(value="VW Owners.xlsx")
-   chat_history = gr.State(value=[])
-   data_visible_state = gr.State(value=False)
-
-   # Title and intro
-   with gr.Column():
-       gr.Markdown("""
-           <h2 style="text-align: center; font-size: 2.25rem; font-weight: 600;">
-               TrustLogic AI Demo
-           </h2>
-       """)
-       gr.Markdown("### Experience TrustLogic AI")
-       gr.Markdown("Experience how quick and easy it is to build the right trust based on insight with our TrustLogic AI")
-
-   # Dataset selection buttons
-   with gr.Column():
-       with gr.Row():
-           vw_customers_btn = gr.Button("VW Owners", elem_id="vw_customers_btn")
-           vw_prospects_btn = gr.Button("VW Prospects", elem_id="vw_prospects_btn")
-
-   # 🧠 Chat section
-   with gr.Column(elem_id="chat_container"):
-       prompt_textbox = gr.Textbox(value=predefined_prompt, scale=4, label="Insert your prompt", visible=True)
-       submit_button = gr.Button("Submit")
-       bot = gr.Chatbot(placeholder=placeholder_text)
-
-   # Button to show data
-   show_data_btn = gr.Button("Show the data", variant="primary", size="lg")
-   
-   # Data analysis section (initially hidden)
-   with gr.Column(visible=False) as data_section:
-       gr.Markdown("### Uncover the underlying data used for the prompt")
-       gr.Markdown("Quickly identify what drives your NPS and trust across different segments using the automated analyser.")
-       gr.Markdown("""
-           <span style="font-size:15px;">Volkswagen Example</span><br>
-           As a default, the analysis displays <strong>Volkswagen Owner</strong> results. 
-           To trigger the analysis for <strong>Prospects</strong>, toggle to 'VW Prospects'.
-       """)
-   
-       with gr.Row(equal_height=True):
-           with gr.Column(scale=1):
-               nps_img_output = gr.HTML()
-   
-           with gr.Column(scale=1):
-               trust_score_output = gr.HTML()
-   
-           with gr.Column(scale=1):
-               gr.Markdown("""<div style='text-align: center;'><h3>How much of your NPS is determined by TrustLogic®</h3></div>""")
-               trust_r2_img = gr.HTML()
-   
-       # Move outputs components inside data_section
-       markdown_3 = gr.Markdown(visible=True)
-       markdown_4 = gr.Markdown(visible=True)
-       trust_plot = gr.Image(value=None, label="Trust Drivers", visible=True)
-       nps_plot = gr.Image(value=None, label="NPS Drivers", visible=True)
-       loyalty_plot = gr.Image(value=None, label="Loyalty Drivers", visible=True)
-       consideration_plot = gr.Image(value=None, label="Consideration Drivers", visible=True)
-       satisfaction_plot = gr.Image(value=None, label="Satisfaction Drivers", visible=True)
-       summary_text = gr.Textbox(value=None, label="Analysis Summary", visible=True)
-       markdown_5 = gr.Markdown(visible=True)
-       markdown_6 = gr.Markdown(visible=True)
-       df_builder_pivot = gr.Dataframe(value=None, label="", visible=True)
-
-   # Create outputs list for functions
-   outputs = [
-       markdown_3, markdown_4, trust_plot, nps_plot, loyalty_plot, 
-       consideration_plot, satisfaction_plot, summary_text, 
-       markdown_5, markdown_6, df_builder_pivot
-   ]
-
-   # Modified process_examples to return values
-   def process_examples_values(file_name):
-       output_components = process_examples(file_name)
-       # Extract values from components
-       values = []
-       for comp in output_components[:11]:  # Only first 11 components
-           if hasattr(comp, 'value'):
-               values.append(comp.value)
-           else:
-               values.append(None)
-       return values
-
-   # Show/hide data section functionality
-   def toggle_data_section(visible):
-       new_visible = not visible
-       button_text = "Hide the data" if new_visible else "Show the data"
-       return gr.update(visible=new_visible), button_text, new_visible
-   
-   show_data_btn.click(
-       fn=toggle_data_section,
-       inputs=data_visible_state,
-       outputs=[data_section, show_data_btn, data_visible_state]
-   )
-   
-   submit_button.click(
-       fn=chatbot_response,
-       inputs=[prompt_textbox, chat_history, selected_dataset_ai],
-       outputs=[bot, chat_history]
-   )
-
-   ## All widget functions here ##
-
-   vw_customers_btn.click(
-       fn=display_trust_score_1,  
-       inputs=[],
-       outputs=trust_score_output,
-   )
-   
-   vw_prospects_btn.click(
-       fn=display_trust_score_2,  
-       inputs=[],
-       outputs=trust_score_output,
-   )
-   
-   vw_customers_btn.click(
-       fn=process_examples_values,
-       inputs=[vw_customers_state],
-       outputs=outputs,
-   )
-   
-   vw_prospects_btn.click(
-       fn=process_examples_values,
-       inputs=[vw_prospects_state],
-       outputs=outputs,
-   )
-   
-   def set_vw_owners():
-       return "VW Owners.xlsx"
-   
-   def set_vw_prospects():
-       return "Volkswagen Non Customers.xlsx"
-
-   vw_customers_btn.click(
-       fn=set_vw_owners,
-       inputs=[],
-       outputs=selected_dataset_ai
-   )
-   
-   vw_prospects_btn.click(
-       fn=set_vw_prospects,
-       inputs=[],
-       outputs=selected_dataset_ai
-   )
-   
-   vw_customers_btn.click(
-       fn=lambda f: vwload_nps_and_r2(f[0]),
-       inputs=[vw_customers_state11],
-       outputs=[nps_img_output, trust_r2_img],
-   )
-   
-   vw_prospects_btn.click(
-       fn=lambda f: load_nps_and_r2(f[0]),
-       inputs=[vw_prospects_state12],
-       outputs=[nps_img_output, trust_r2_img],
-   )
-           
-   demo.load(
-       fn=lambda f: vwload_nps_and_r2(f[0]),
-       inputs=[vw_customers_state11],
-       outputs=[nps_img_output, trust_r2_img],
-   )
-   
-   demo.load(
-       fn=display_trust_score_1, 
-       inputs=[], 
-       outputs=trust_score_output
-   )
-   
-   demo.load(
-       fn=process_examples_values, 
-       inputs=[vw_customers_state], 
-       outputs=outputs
-   )
+    gr.HTML("""
+    <style>
+    #trust_driver_img, #nps_driver_img {
+        display: inline-block;
+        width: 49%;
+        margin-right: 1%;
+        vertical-align: top;
+    }
+    </style>
+    """)
+
+    # Title and intro
+    with gr.Column():
+        gr.Markdown("""
+            <h2 style="text-align: center; font-size: 2.25rem; font-weight: 600;">
+                What drives your NPS and trust?
+            </h2>
+        """)
+        gr.Markdown("Quickly identify what drives your NPS and trust across different segments using the automated analyser.")
+        gr.Markdown("""
+            <span style="font-size:15px;">Volkswagen Example</span><br>
+            As a default, the analysis displays <strong>Volkswagen Owner</strong> results. 
+            To trigger the analysis for <strong>Prospects</strong>, toggle to ‘VW Prospects’.
+        """)
+
+    with gr.Column():
+        with gr.Row():
+            vw_customers_btn = gr.Button("VW Owners", elem_id="vw_customers_btn")
+            vw_prospects_btn = gr.Button("VW Prospects", elem_id="vw_prospects_btn")
+    
+        with gr.Row(equal_height=True):
+            with gr.Column(scale=1):
+                nps_img_output = gr.HTML()
+
+                
+    
+            with gr.Column(scale=1):
+                trust_score_output = gr.HTML()
+    
+            with gr.Column(scale=1):
+                gr.Markdown("""<div style='text-align: center;'><h3>How much of your NPS is determined by TrustLogic®</h3></div>""")
+                trust_r2_img = gr.HTML()
+    
+                
+        with gr.Column():
+            outputs = reset_outputs()
+
+        # 🔁 States
+        vw_customers_state11 = gr.State(value=["example_files/VW Owners.xlsx"])
+        vw_prospects_state12 = gr.State(value=["example_files/Volkswagen Non Customers.xlsx"])
+        vw_customers_state = gr.State(value=["VW Owners.xlsx"])
+        vw_prospects_state = gr.State(value=["Volkswagen Non Customers.xlsx"])
+        selected_dataset_ai = gr.State(value="VW Owners.xlsx")  # ✅ Matches dictionary key
+
+        chat_history = gr.State(value=[])
+
+    # 🧠 Chat section
+    with gr.Column(elem_id="chat_container"):
+        gr.Markdown("###  Test-drive the results in the TrustLogicAI")
+        gr.Markdown("Our AI uses the analysis results to generate trust-optimised content.")
 
+        prompt_textbox = gr.Textbox(value=predefined_prompt, scale=4, label="Insert your prompt", visible=True)
+        submit_button = gr.Button("Submit")
+        bot = gr.Chatbot(placeholder=placeholder_text)
+
+        
+        
+        submit_button.click(
+            fn=chatbot_response,
+            inputs=[prompt_textbox, chat_history, selected_dataset_ai],
+            outputs=[bot, chat_history]
+        )
+
+
+
+        ## All widget functions here ##
+
+        vw_customers_btn.click(
+            fn=display_trust_score_1,  
+            inputs=[],
+            outputs=trust_score_output,
+        )
+        
+        vw_prospects_btn.click(
+            fn=display_trust_score_2,  
+            inputs=[],
+            outputs=trust_score_output,
+        )
+        
+        vw_customers_btn.click(
+            fn=process_examples,
+            inputs=[vw_customers_state],
+            outputs= outputs,
+        )
+        
+        vw_prospects_btn.click(
+            fn=process_examples,
+            inputs=[vw_prospects_state],
+            outputs= outputs,
+        )
+        def set_vw_owners():
+            return "VW Owners.xlsx"
+        
+        def set_vw_prospects():
+            return "Volkswagen Non Customers.xlsx"
+
+        
+        vw_customers_btn.click(
+            fn=set_vw_owners,
+            inputs=[],
+            outputs=selected_dataset_ai
+        )
+        
+        vw_prospects_btn.click(
+            fn=set_vw_prospects,
+            inputs=[],
+            outputs=selected_dataset_ai
+        )
+        vw_customers_btn.click(
+            fn=lambda f: vwload_nps_and_r2(f[0]),
+            inputs=[vw_customers_state11],
+            outputs=[nps_img_output, trust_r2_img],
+        )
+        
+        vw_prospects_btn.click(
+            fn=lambda f: load_nps_and_r2(f[0]),
+            inputs=[vw_prospects_state12],
+            outputs=[nps_img_output, trust_r2_img],
+        )
+                
+        
+        demo.load(
+            fn=lambda f: vwload_nps_and_r2(f[0]),
+            inputs=[vw_customers_state11],
+            outputs=[nps_img_output, trust_r2_img],
+        )
+        demo.load(
+        fn=display_trust_score_1, 
+        inputs=[], 
+        outputs=trust_score_output
+        )
+        
+        demo.load(
+            fn=process_examples, 
+            inputs=[vw_customers_state], 
+            outputs=outputs
+        )
+        
 try:
-   demo.launch(server_name="0.0.0.0")
+    demo.launch(server_name="0.0.0.0")
 except Exception as e:
-   logger.error(f"Error launching Gradio app: {e}")
-   raise e
-
+    logger.error(f"Error launching Gradio app: {e}")
+    raise e